Cluster Analysis with Grey Comprehensive Correlation Coefficient for Prediction of Bridge Damage

In this paper, the grey comprehensive correlation coefficient is used to determine the weights for cluster analysis. A full-link network model of distresses of bridge structures and bridge deck pavement is established, and the grey comprehensive correlation coefficients between each distress type are obtained. The cluster analysis method is used to group the bridge deck pavement distresses. The structural damage trend was estimated with a feedback prediction, and some suggestions were proposed to avoid the risk of serious structural damage. Through calculation analysis and verification of the corresponding test data, it can be concluded that the weighted clustering method based on gray correlation weighting is feasible for the correlation analysis and trend estimation of bridge distress types. This method can help do pre-judgment for hidden structural damage, and provide a reference for the next step to establish a structural damage process prediction model and bridge intelligent management.